Qualification mismatches typically come from parsing or normalization gaps between the JD and the resume. Use this checklist:
1) Verify parsed data exists and is comparable
- Open the parsed JD JSON and confirm the required fields for education/qualification (e.g., degree, fields of study) are present.
- Open the parsed resume JSON and confirm education entries (degree, major, level) are extracted—watch for image-only PDFs or poor OCR.
2) Check normalization & taxonomy
- Confirm both JD and resume degrees are normalized to the same education taxonomy (e.g., “B.Tech” ↔ “Bachelor of Technology”).
- If your use case uses localized terms (BSc/HND/Licenciatura), ensure language/locale and taxonomy version match your environment.
- Add custom synonyms/aliases for common degree variants to improve alignment.
3) Review OneMatch configuration
- Validate weights/thresholds for education vs. other factors (skills, titles).
- Check any must-have education rules and ensure they correspond to normalized degree names (not raw text).
- Confirm the same taxonomy/synonym set and parser version are used across runs.
4) Diff old vs. new results
- Compare previous vs. current OneMatch outputs side-by-side focusing on education/qualification nodes to see what changed (missing fields, different degree labels, locale shifts).
Share the affected JD and resume files plus their parsed JSON and your OneMatch config (weights/taxonomy settings) with support@rchilli.com for a deep-dive and targeted fixes.
Comments
0 comments
Please sign in to leave a comment.